BBA - Bioenergetics 1860 (2019) 708–716

Contents lists available at ScienceDirect

BBA - Bioenergetics

journal homepage: www.elsevier.com/locate/bbabio

Tryptophan 224 of the rat mitochondrial carnitine/acylcarnitine carrier is T crucial for the antiport mechanism Nicola Giangregorioa,1, Annamaria Tonazzia,1, Lara Consoleb, Mariella Pistilloc, Vito Scalerac, ⁎ Cesare Indiverib, a CNR Institute of Biomembranes, Bioenergetics and Molecular Biotechnology, via Amendola 165/A, 70126 Bari, Italy b Department DiBEST (Biologia, Ecologia, Scienze della Terra) Unit of Biochemistry and Molecular Biotechnology, University of Calabria, Via P. Bucci 4C, 87036 Arcavacata di Rende, Italy c Department of Biosciences, Biotechnology and Biopharmaceutics, University of Bari, via Orabona 4, 70126 Bari, Italy

ARTICLE INFO ABSTRACT

Keywords: The mitochondrial carnitine/acylcarnitine carrier (CACT) catalyzes an antiport of carnitine and acylcarnitines Carnitine and also a uniport reaction with a rate of about one tenth with respect to the antiport rate. The antiport process Mitochondria results from the coupling of the two uniport reactions in opposite directions. In this mechanism, the transition of Membrane transport the carrier from the outward open conformation to the inward open one (or vice versa) is much faster for the Site-directed mutagenesis carrier-substrate complex than for the unbound carrier. To investigate the molecular determinants that couple the binding of the substrate with the conformational transitions, site directed mutagenesis has been employed. The antiport or the uniport reaction was followed as [3H]carnitine uptake in or efflux from proteoliposomes reconstituted with the WT or Trp mutants of the rat CACT. Substitution of each the three Trp residues led to different results. Nearly no variations were observed upon substitution of W192 and/or W296 with Ala.While, substantial alteration of the transport function was observed in the mutants W224A, W224Y and W224F. Mutation of W224 led to the loss of the antiport function while the uniport function was unaltered. In these mutants impairment of the substrate affinity on the external side was also observed. The data highlights that W224 is involved in the coupling of the substrate binding with the matrix gate opening. The experimental data are in line with predictions by homology modeling of the CACT in its cytosolic (c-state) or matrix (m-state) opened conformations.

1. Introduction facing the matrix, connect the transmembrane segments within each couple. The substrate-binding sites are located in the central cavity at The mitochondrial carriers, belonging to the SLC25 family, share about mid of the membrane [3–8]. The recent structure of the ANC peculiar structural features and molecular mechanisms [1] which have from T. thermophila in complex with the bongkrekic acid (PDB code been firstly revealed by the 3D structure of the Adenine nucleotide- 6GCI), in the matrix-opened conformation (m-state), added important carrier (ANC) in a conformation opened towards the cytosol (c-state) information on the molecular mechanism of the transport reaction [2]. The structures of the other members of the SLC25 family have been catalyzed by mitochondrial carriers. The SLC25 predicted by homology modeling, using the ANC 3D coordinates as the family counts many members each specific for few substrates. These template (PDB code 1OKC). A bundle of six α-helix transmembrane transporters catalyze the exchange of many metabolites from the cy- segments, enclosing a water-filled cavity, characterizes the carrier tosol to the mitochondrial matrix and vice-versa allowing completion of structures. Repetitions of two adjacent transmembrane seg- metabolic pathways which are characterized by the localization of en- ments show internal similarity highlighting a tripartite structure. Short zymes into both compartments. To perform this task most mitochon- hydrophilic loops with random coiled structure, facing the inter- drial carriers catalyze antiport reactions. In this transport mode, the membrane space, connect the hydrophobic transmembrane segment transport of one substrate from the cytosol to the matrix requires the repetitions while, larger hydrophilic loops with a α-helical structure, counter-transport of another substrate in the opposite direction. Two

⁎ Corresponding author. E-mail address: [email protected] (C. Indiveri). 1 These authors contributed equally to this work. https://doi.org/10.1016/j.bbabio.2019.07.006 Received 31 March 2019; Received in revised form 6 July 2019; Accepted 18 July 2019 Available online 21 July 2019 0005-2728/ © 2019 Elsevier B.V. All rights reserved. N. Giangregorio, et al. BBA - Bioenergetics 1860 (2019) 708–716 networks of six amino acid residues present at the bottom (close to the previously described [26] was amplified using the forward primer ta- matrix) and at the top of the transporters (close to the intermembrane tatcatatgtccgccgccgaatcgccc and the reverse primer gaacggatccttagtc- space) play a crucial role in the molecular mechanism of the antiport. gaaaagcttgttcatga, containing the Nde I and BamHI restriction sites, The subsequent opening and closing of these networks together with respectively, for cloning into the pMW7 vector. The purified cDNA was the simultaneous rotation of the three intermembrane domains, allows cloned in pMW7 for transforming E. coli C0214 and over-expressing the substrates to enter the carrier from one side and to be translocated protein. Isolation, solubilization and purification of the wild-type AcuH towards the opposite side. The occurrence of this process in the two were performed as described above for the rat CACT [23,25]. To in- opposite directions realizes the alternating access mechanism, which is troduce specific mutations in the AcuH the same procedure at the basis of the antiport function. Antiport is favored with respect to used for the rat CACT was used [23,24]. uniport, since the gate opening is much faster for the carrier-substrate complex than for the empty carrier [9]. Several information on the 2.3. Reconstitution of wild-type (WT) and mutant proteins in liposomes structure/function relationships has been obtained by site-directed mutagenesis [1,4,9–15]. However, the molecular determinants re- The recombinant proteins (60 μg protein) were reconstituted into sponsible for coupling substrate binding with gate opening that favors liposomes as described previously [7,23,25]. The concentration of in- the antiport, are still unknown. Some information derives from studies traliposomal carnitine was 15 mM except when differently specified in on the carnitine/acylcarnitine carrier (CACT) that has the peculiarity, the legends to figures. The orientation of the reconstituted protein was among mitochondrial carrier, of catalyzing a measurable uniport re- the same as in the native membrane as previously described [16]. action that occurs at a lower rate with respect to the antiport reaction. This special feature could allow discriminating the molecular de- 2.4. Transport measurements terminants crucial for the sole antiport reaction from those important for the overall mechanism of catalysis. Previous data highlighted that In order to remove the external substrate for antiport or uniport an amino acid residue belonging to the matrix network, namely K35, is assay, 550 μ1 proteoliposomes obtained as in Section 2.3, were passed involved in the antiport mechanism [16]. Trp residues have been shown through a Sephadex G-75 column (0.7 cm × 15 cm), equilibrated with in bacterial antiporters, as well as in some human orthologues, to play 10 mM PIPES pH 7.0 and 90 mM NaCl. The eluted proteoliposomes important roles in transport mechanism both in gating and in substrate (600 μl), distributed in reaction vessels (100 μl), were used for transport recognition [17–21]. Thus, we have now investigated the role of a Trp measurements using the inhibitor-stop method [27]. Transport was residue, which is conserved in the CACT sub-family and is rarely pre- performed at 25 °C, started by adding 0.1 mM [3H]carnitine (except sent in other mitochondrial carriers. Some crucial experiments have that in kinetic experiments in which different [3H]carnitine con- also been performed on the A. nidulans CACT which, differently from centrations were used) to proteoliposomes and terminated by the ad- the rat transporter, contains the sole conserved Trp residue. dition of 1.5 mM NEM (N-ethylmaleimide), a specific inhibitor of CACT. In control samples, the inhibitor was added together with the radi- 2. Materials and methods olabeled substrate at time zero. After stopping the transport reaction, the radioactivity not taken up by the proteoliposomes was removed by 2.1. Materials passing 100 μl of each sample through a Sephadex G-75 column (0.6 cm × 8 cm). The proteoliposomes were eluted with 1.2 ml 50 mM Sephadex G-75 was purchased from Pharmacia, egg-yolk phospho- NaC1 and collected in 4 ml scintillation mixture, vortexed and counted. lipids (l-α-phosphatidylcholine from fresh turkey egg yolk) and The experimental values were corrected by subtracting control values. Amberlite XAD-4 from Fluka, PIPES, Triton X-100, cardiolipin, N-do- For kinetic experiments, initial transport rate was derived from the decanoylsarcosine (sarcosyl), Koshland's reagent (KR) from Sigma, St. initial linear part of the time course of the isotope equilibration ki- Louis, MO, L-[methyl-3H]carnitine was from Amersham. All reagents netics. For the purpose of performing efflux experiments, proteolipo- were of analytical grade. somes were filled with the indicated concentration of[3H]carnitine (intraliposomal carnitine) before starting efflux transport in uniport 2.2. Over-expression and isolation of the CACT proteins mode or antiport mode [16]. NEM (1.5 mM) was added for stopping the transport reaction at various times. The decrease in radioactivity inside The previously constructed pMW7-WT rat CACT recombinant the proteoliposomes was measured and expressed as percent of the plasmid [22] was used to introduce specific mutations (see Results) in radioactivity present at time zero [16,27]. the CACT protein as previously described [23]. The amino acid re- placements were performed with complementary mutagenic primers 3. Results using the overlap extension method [24] and the High Fidelity PCR System (Roche). The PCR products were purified by the QIAEX II Gel 3.1. Location of the Trp residues in the CACT homology structural model Extraction Kit (QIAGEN La Jolla), digested with NdeI and HindIII (re- striction sites added at the 5′ end of forward and reverse primers, re- The rat CACT contains three Trp residues, namely W192, W224 and spectively) and ligated into the pMW7 vector that was previously di- W296. W224 is conserved along with the different species of the CACT gested with the same restriction enzymes. Sequencing and subfamily while the others are present only in the vertebrate CACTs transformation of the selected clones into Escherichia coli C0214, bac- (Fig. 1). The predicted positions of the Trp residues are depicted in the terial over-expression, isolation of the inclusion body fraction, solubi- homology structural model of CACT by licorice representation (Fig. 2). lization and purification of the wild-type CACT and mutant CACT W224 is located in the central aqueous cavity of the carrier, in the vi- proteins were performed as described previously [23,25]. cinity of other residues that constitute the matrix gate of the transporter In order to over-express the fungus Aspergillus nidulans carnitine/ [8]. W192 is located towards the cytosolic side of the transporter acylcarnitine carrier (AcuH) protein, the CACT cDNA obtained as (Fig. 2). W296 is also located in the cytosolic side, but is not depicted in

709 N. Giangregorio, et al. BBA - Bioenergetics 1860 (2019) 708–716

Fig. 1. Alignment of proteins of the mitochondrial carnitine carrier subfamily. CACT proteins (NP446417 from Rattus norvegicus, NP000378 from Homo sapiens, NP568670 from Arabidopsis thaliana, NP014743 from Saccharomyces cerevisiae, AJ011563 from Aspergillus nidulans, NP065266 from Mus musculus, XP001253588 from Bos taurus, NP957153 from Danio rerio, NP501223 from Caenorhabditis elegans, NP477221 from Drosophila melanogaster, and NP001065471 from Oryza sativa) were aligned using the Clustal Omega software. Identities are indicated by asterisks and conservative or highly conservative substitutions are indicated by dots or colons, respectively. The three Trp residues 192, 224 and 296 are highlighted by dark background.

Fig. 2. Location of the CACT Trp residues in the homology structural model. The homology model has been obtained by Swiss model portal using as templates the 1OKC PDB and represented by the molecular visualization program VMD. A) Lateral view of CACT in c-state depicted as a ribbon diagram. Residues of the matrix gate are highlighted in red with a transparent space filed conformation, the residues of the central binding site are depicted in green with a ball and stick diagram while Tryptophans are highlighted in blue as sticks graphic representation. B) Top view of the same CACT structure.

710 N. Giangregorio, et al. BBA - Bioenergetics 1860 (2019) 708–716

The low activity of the mutants lacking W224 suggests a crucial role for this residue in transport function. To gain further insights into the re- lationships among W224 and the transport mechanism, the analysis of the kinetic parameters of the mutant CACTs were performed (Table 1) in comparison to those of the WT. Km values more than eight times that of WT were measured for the mutants lacking W224. While, the mu- tants W192A, W296A and W192/296A displayed a Km value similar to that of the WT. The Vmax was also impaired in the W224A, W-less, W224F and W224Y mutants, while, the W192A, W296A and the double W192/296A mutants showed Vmax similar or not very different from that of the WT (Table 1). The catalytic efficiencies of the mutants were calculated and compared with that of the WT. The impairments of catalytic efficiencies reflected the impairments in Km and Vmax.The strong increases in Km (decreased affinity) of the W224 mutants in- 3 Fig. 3. Time course of [ H]-carnitine uptake in liposomes reconstituted with dicates that this residue could be crucial in the interaction with the recombinant WT and Trp mutant proteins of CAC. Transport was started by the substrate in some steps of the transport process. addition of 0.1 mM [3H]-carnitine to proteoliposomes containing 15 mM car- nitine and stopped at the indicated times. The plot shows the time course curves 3.3. Inhibition of the WT and CACT mutants by the Koshland's reagent of WT and mutant CACT as reported in the figure legend. The data represent means ± SD of at least three independent experiments. Chemical modification of protein can give additional information for identifying amino acid residues involved in the binding of the Table 1 substrate [8,28,29]. Koshland's reagent (KR) specifically reacts with Kinetic analysis of WT and mutant CACTs. Kinetic analysis was performed as tryptophan residues and, then, could be employed for proofing the in- described in Materials and methods. Initial transport rates were measured at volvement of Trp residues in the interaction with the substrate [30,31]. various substrate concentrations as shown for Fig. 4 B and interpolated in the Inhibition by this reagent was tested on the transport activity of WT Lineweaver-Burk equation for obtaining the Km and Vmax values. The reported CACT and on some representative mutants. Both the WT and the W192/ data are means ± S.D. from three independent experiments. 296A mutants (Fig. 4A) displayed strong inhibition by the reagent. Protein External Vmax Vmax/Km From the dose-dependence analysis, half inhibition constants (IC50) Km (mM) (μmol/min/mg) (ml/min/mg) were derived. The calculated values, 2.7 ± 0.35 mM for the WT and

WT 0.40 ± 0.10 1.3 ± 0.7 3.3 2.9 ± 0.25 mM for the W192/296A, are very similar and confirm that W192A 0.66 ± 0.014 1.6 ± 0.071 2.4 the two proteins share the same response to the modification of Trp W224A 3.7 ± 1.5 0.42 ± 0.19 0.11 residue(s). As expected, the mutant W-less was insensitive to the re- W224F 3.2 ± 0.86 0.38 ± 0.068 0.12 agent. The mutant W224A was also insensitive as the W-less, indicating W224Y 3.3 ± 1.0 0.31 ± 0.12 0.10 that the sole modification of W224 is responsible for the activity loss W296A 0.31 ± 0.0071 0.61 ± 0.12 2.0 W192/296A 0.47 ± 0.080 1.4 ± 0.7 3.0 caused by the reagent. However, it cannot be excluded that W192 and/ W-less 4.4 ± 0.52 0.26 ± 0.16 0.060 or W296 could be modified by the reagent, but, in this case, the reaction has no consequences on the transport function, indicating that the re- sidues are not critical for function and are located in protein moieties the CACT structural model, since this structure derives from that of the not involved in the transport path, according to the structure (Fig. 2). ADP/ATP carrier that lacks the last C-terminal amino acids. Very interestingly, the Koshland's reagent did not have any effect on A role of W224 in substrate binding and translocation was expected, WT and on the W292/296A double mutant when the uniport activity since this residue is located in the core of the protein, close to the was measured instead of antiport. This data suggests that the sole an- matrix gate and to the substrate binding site. The Trp residues were tiport function is sensitive to the chemical modification of W224. The substituted by Ala, generating the single mutants W192A, W224A, and single mutants W192A and W296A showed the same behavior as the W296A. Since W224A was the sole mutant showing impaired function WT (not shown) according to the general features of these mutants (see (see below), additional mutants were constructed to better understand also Table 1). To investigate the relationships between W224 and the the role of W224, i.e., the double mutant W192A/296A, and the triple substrate binding site, inhibition kinetics were performed in the pre- mutant W-less. W224 was also substituted by the conservative amino sence or absence of the Koshland's reagent and the data was analyzed acids Phe and Tyr, generating the W224F and W224Y mutants. The by a double reciprocal plot, as shown in Fig. 4B. The data points in- recombinant mutant proteins were over-expressed in E. coli and re- terpolate straight lines intersecting on the X axis, indicating a non- constituted in liposomes as described in Material and methods section competitive inhibition. The half saturation constants (Ki) calculated [23]. The amount of each mutant protein incorporated into proteoli- from the graph were 8.5 ± 1.3 or 5.9 ± 1.0 mM for the WT or the posomes was very similar to that of the WT (not shown) and ranged mutant W192/296A, respectively. This pattern suggests, apparently, between 25 and 35% of the recombinant protein added for recon- that the inhibitor binding site could be far from the substrate binding stitution, as found for many other previously constructed mutants of site. However, in the case of covalent inhibition, this type of pattern can CACT [8,16,23]. occur even if the reagent target is close to the substrate binding site [32]. Thus, a protection analysis was performed, that in such cases may 3.2. Transport activity of the Trp mutants of CACT reveal the interaction of the inhibitor and the substrate with the same (or vicinal) sites. To investigate the possible protective effect of carni- The transport activities of WT and Trp mutants were measured as a tine on the binding of the Koshland's reagent to the sole W224, the function of time. Fig. 3 shows that W192A, W296A and the double proteoliposomes reconstituted with the double mutant W192/296A mutant W192/296A (containing only W224) exhibited transport ac- (containing the sole W224) were incubated with the inhibitor in the tivities comparable to that of WT; whereas the single mutants W224A, presence or absence of carnitine. Koshland's reagent concentration was W224Y and the W-less showed much lower transport activities, with kept close to its Ki. As shown in Fig. 4C, the mutant exhibited sub- transport at 60 min of about 15% of WT. The mutant W224F showed a stantial protection by the substrate. Indeed, the 70% inhibition of the transport activity slightly higher than the other mutants lacking W224. control, without carnitine pre-incubation, decreased to about 50% and

711 N. Giangregorio, et al. BBA - Bioenergetics 1860 (2019) 708–716

38% upon pre-incubation with 0.5 or 10 mM carnitine, respectively. This data correlates well with the vicinity of the W224 to the residues of the matrix network and to the residues of the substrate binding site (Fig. 2). In the basis of the described data, it seems that the substitution of W224 impairs the coupling of the substrate binding with the matrix network opening, abolishing the antiport function. This hypothesis correlates well with the data of Fig. 3, in which the transport activity of the mutant W224 is about 10% that of WT indicating that the impaired function could be only due to the absence of stimulation by the counter substrate with the sole remaining uniport function.

3.4. Loss of antiport function of the W224 CACT mutants

To gain further insights into the impairment of the antiport func- Fig. 5. Stimulation of carnitine transport by internal countersubstrate. tion, the stimulation by the countersubstrate was tested by measuring Transport was started by the addition of 0.1 mM [3H]-carnitine to proteolipo- 3 the transport activity as [ H]carnitine transport in the presence of in- somes containing the indicated carnitine concentrations and stopped at 10 min. creasing internal substrate concentrations (Fig. 5). As expected for the The data related to the WT or W224A CACT proteins are shown as white or gray WT, transport increased by increasing the intraliposomal carnitine bars, respectively. The data represent means ± SD of three independent ex- concentration according to the antiport mode. While, in the case of the periments. (*), Significantly different from the control (without intraliposomal W224A mutant, no variations of [3H]carnitine uptake were measured, carnitine) for p < 0.05, as calculated from Student's t-test analysis.

Fig. 4. Inhibition of the CACT by Koshland's reagent. (A), Dose-response curve for the inhibition of the reconstituted recombinant WT protein and Trp mutants of CACT. The antiport rate was measured adding 0.1 mM [3H]carnitine to proteoliposomes containing 15 mM internal carnitine. The uniport rate was measured adding the labeled carnitine to proteoliposomes without internal substrate. KR, at the indicated concentrations, was added 2 min before the transport assay. Percent of residual activities are reported. The values are means ± S.D. from three independent experiments. (B), Kinetic analysis of the effect of Koshland's reagent on theWT and W192/296A proteins. The antiport rate was measured adding [3H]carnitine at the indicated concentrations to proteoliposomes containing 15 mM internal carnitine, in the absence (○, □) or in the presence of 3.0 mM KR added together with the labeled substrate (●, ■). Experimental data are plotted according to Lineweaver-Burk as reciprocal transport rate vs reciprocal carnitine concentrations. Reported values are means ± S.D. from three independent experiments. (C), substrate protection of the W192/296A CACT inhibition by KR. Carnitine at the indicated concentrations (0.5 and 10 mM) was added to proteoliposomes passed through a Sephadex G-75 column and incubated for 5 min with KR 3.0 mM. Then, the unreacted compound was removed by passing the proteoliposomes through a second Sephadex G-75 column. Transport was started by adding 0.1 mM [3H]carnitine to the proteoliposomes and stopped after 30 min. Percent residual activity is reported with respect to the control without the addition of the inhibitor. Reported values were the means ± SD from three different experiments. (*), Significantly different from the control (without added carnitine) for p < 0.05, as calculated from Student's t-test analysis.

712 N. Giangregorio, et al. BBA - Bioenergetics 1860 (2019) 708–716

Fig. 6. Antiport and uniport function of WT and mutant CACTs. To measure the transport activity of CACT using the efflux approach, proteoliposomes reconstituted with WT or W mutants, were preloaded3 with15mM[ H] carnitine, then the efflux of labeled carnitine was monitored along time. The residual intraliposomal radioactivity was measured at the indicated timesinthe presence of 5 mM external carnitine (antiport) or buffer alone (uniport). The proteins (WT or mutants tested) in A, B and C are indicated in the figures. Thevaluesare means ± SD from three experiments. i.e., no stimulation by the countersubstrate could be detected, con- values were obtained, i.e., 7.3 ± 2.8 mM or 6.7 ± 3.0 mM (from 3 firming that this mutant catalyzes only the residual uniport function. independent experiments) for the WT and the W224A, respectively, For a final proof of this hypothesis, the efflux procedure wasem- suggesting that W224 is not involved in substrate binding from the ployed which is the most reliable to discriminate between antiport and internal side. uniport function (Fig. 6). To this aim, [3H]carnitine efflux from pre- To further support the specific role of W224 in the antiport me- loaded proteoliposomes was measured in the absence (uniport) or chanism, the same mutation was performed on the A. nidulans CACT, presence (antiport) of the external substrate. The WT proteoliposomes that is quite different from the human protein, but has the sole W224 behaved as expected, i.e., antiport was much more efficient than uni- homologous residue, W254. The antiport and uniport reactions were port. Similarly, the mutants in which one of the two non-critical Trp measured in proteoliposomes by the efflux procedure (Fig. 7). Also in residues were substituted, showed an antiport function much more ef- this case, the antiport function virtually disappeared in the mutant ficient than the uniport. Again, a similar result was obtained withthe W254A, confirming that this residue is crucial for the antiport me- double mutant W192/296A (Fig. 6B). On the contrary, the W224A and chanism. the W-less did not show stimulation by external substrate, demon- strating that the antiport function was impaired. Slight stimulation of the antiport by the addition of the external substrate was only observed 4. Discussion with the W224F (Fig. 6 C). The efflux procedure was also used to measure the internal Km for W224 in comparison with the WT. Similar The molecular determinants of the transport mechanism of SLC members are little known. More information is available for bacterial

713 N. Giangregorio, et al. BBA - Bioenergetics 1860 (2019) 708–716

carnitine that, together with K35, facilitates the opening of the trans- porter towards the matrix. In the present work, the second, most crucial residue has been identified, namely W224. Indeed, the substitution of W224 completely abolishes the antiport, leaving the sole uniport function. The protein lacking W224 is thus converted into a uniporter with a transport activity equal to the unidirectional activity of the WT. Indeed, the effect of mutation of W224 on antiport impairment ismuch more evident than in the case of substitution of K35 [16]. Moreover, the Koshland's reagent that reacts with Trp residues, affect only the antiport function, but not the uniport, confirming that W224 is crucial for an- tiport. Altogether, the results indicate that W224 is the major residue involved in coupling and it undergoes to a cation-π bond with the ammonium moiety of carnitine. It is well known than cation-π is the strongest among the non-covalent interactions [37]. Indeed, betaine, which shares a quaternary ammonium moiety with carnitine, binds to Trp residues in the bacterial Glycine Betaine Transporter [37,38]. In CACT, the interaction of carnitine with the two residues W224 and K35, the second of which is in the matrix gate [16], induces a fast opening of Fig. 7. Antiport and uniport function of WT and W245A mutant A. nidulans the gate. In the absence of the substrate, the gate opening is much CACT. slower. The distance of W224 from K35 (3 Å) corresponds to the dis- To measure the transport activity of A. nidulans CACT using the efflux approach, tance between the carboxylic and the ammonium groups of carnitine proteoliposomes reconstituted with WT or W245A mutant, were preloaded with (Fig. 8A and B). Indeed, the substitution of W224 impairs this catalytic 15 mM [3H]carnitine, then the efflux of labeled carnitine was monitored along step and the opening process is slowed down, matching the rate of time. The residual intraliposomal radioactivity was measured at the indicated transition of the unbound carrier. The substitution overlaps the effect of times in the presence of 5 mM of external carnitine (antiport) or buffer alone chemical modification [30] of the W224, leading to the same results. (uniport). The values are means ± SD from three experiments. The coupling of substrate binding with gate opening occurs when the substrate enters the carrier from the external side (Fig. 8A and B). This [17,20] than for eukaryotic transporters. Concerning the mitochondrial correlates well with the increase in external Km of the proteins lacking carriers, these transport proteins can be assimilated to single binding W224 with respect to the WT. While, W224 is not important for the center gated pores, as some bacterial transporters. However, mi- substrate binding and transport from inside to outside, i.e., when the tochondrial carriers are unique with respect to their structure and, matrix gate is opened. In this state the distance between W224 and K35 hence, mechanism since these proteins are constituted by a bundle of is greater than in the c-state and, hence, the two residues are not sui- six transmembrane segments arranged in three intramembrane domains table for binding of the internal substrate (Fig. 8C and D). This again which rotate to allow conformational changes for transport reaction correlates with the lack of changes in the internal Km upon substitution [11,33]. The structures of the bovine and yeast ADP/ATP carriers were of the W224. The experimental data correlates very well with the firstly available in the c-state conformation opened towards the cyto- change of the reciprocal position of the two residues in the structures of solic side [2,34]. The structure of the ADP/ATP carrier from T. ther- the CACT in the c-state and in the m-state (Fig. 8) obtained by mophila is now available in the m-state, as well [9]. This last structure homology modeling on the basis of the two structures of the ADP/ATP led to significant advancements in defining the conformational changes carrier, respectively [2,9]. The data show that the antiport is impaired required for substrate transfer from one side of the membrane to the in both the directions since the lack of one of the fast steps (from other. This information together with in-silico analysis and site-directed outside to inside) limits the overall rate of conformational transitions. mutagenesis studies, indicate that substrate binding induces the con- The binding is only impaired from the external (higher Km) since only formational changes required for transport [9]. The antiport me- the residue involved in binding from the external side is substituted. chanism is determined by the coupling of substrate binding with gate Hence, the residue(s) involved in binding and gate opening from the opening. Indeed, the rate of transition of the carrier from an outward internal side, if present, should be different and probably located in the open conformation (c-state) to the inward open one (m-state) or vice vicinity of the cytosolic gate residues [5]. versa, is forbidden or much slower for the unbound carrier than for the The data are also in line with the described interaction of the CACT carrier-substrate complex [33]. with omeprazole. Interestingly, W224, as well as K35 are among the In this work, we provide further insights into the mechanism of residues interacting with omeprazole molecules docked into the CACT coupling of substrate binding with the opening of the matrix gate of the structure. Given that omeprazole is not transported, the inhibitor may CACT, which is at the basis of the antiport mechanism. This transporter bind the two residues involved in carrier opening without inducing is suitable to gain insights into the molecular mechanism of the antiport conformational changes, but blocking the carrier in a non-functional since it is among the few transporters that, besides antiport, also cat- conformation. The occurrence of many other interactions with ome- alyzes a uniport reaction with a lower rate than the antiport. This pe- prazole prevents the carrier opening [39]. culiarity has been proposed since the 80s in studies performed with W224 is present also in the Ornithine/Citrulline carrier isoforms intact mitochondria [35] and then confirmed by detailed studies per- (SLC25A15/A2) and in the mitochondrial transporter for basic amino formed with both the native protein from rat liver and the recombinant acids (SLC25A29) [40]. Indeed, the substrates of these transporters rat and human proteins [16,36]. The co-existence of the two modes show some similarity with carnitine, i.e., a carboxylic as well as a allows discriminating mutations, which impair the overall transport charged amino group at a distance of 3–4 Å. Interestingly, the pre- cycle from those, which specifically impair antiport. Using this strategy, viously reported mutation of W224 with Ala in the ORNT1 (SLC25A15), we identified an amino acid residue, namely K35, whose substitution causes a strong reduction of antiport of Ornithine and Arginine, while caused a reduction of transport stimulation by counter-substrate, im- the very slow unidirectional transport (with sodium inside the proteo- pairing the antiport function. Those findings suggested that K35 should liposomes) remains about the same of the WT protein [41]. Moreover, be involved in the binding of the carboxyl moiety of carnitine favoring also the SLC25A29 shows a uniport function with a lower efficiency the gate opening. However, at least one companion amino acid residue with respect to the antiport function similarly to the CACT [40]. The should then be involved in the binding of the ammonium moiety of lack of this Trp residue in other carriers indicates that the highlighted

714 N. Giangregorio, et al. BBA - Bioenergetics 1860 (2019) 708–716

Fig. 8. Amino acid residues of CACT involved in the coupling of substrate binding with gate opening. The homology models of CACT in c-state and m-state have been obtained by Swiss model portal using as templates the 1OKC or 6GCI PDB, respectively and represented by the molecular visualization program VMD, as indicated in the Fig. 2 legend. Carnitine docking analysis has been performed using ArgusLab software. A) Lateral view of CACT structure in c-state with bound carnitine was highlighted as a ribbon diagram. Trp 224 and K35 were in blue and red re- spectively. Carnitine was depicted with a licorice conformation. B) Top view of the same CACT struc- ture. C) Lateral view of CACT structure in m-state highlighted as a ribbon diagram. Trp 224 and K35 were in blue and red respectively. D) Bottom view of the same CACT structure.

mechanism might be typical of transporters handling substrates with a carriers, Biochim. Biophys. Acta 1757 (2006) 1237–1248. positively charged functional group. Carriers handling anions must [4] A.J. Robinson, E.R. Kunji, Mitochondrial carriers in the cytoplasmic state have a common substrate binding site, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) have different mechanisms or different type of amino acid residues 2617–2622. involved in similar mechanisms. [5] A.J. Robinson, C. Overy, E.R. Kunji, The mechanism of transport by mitochondrial carriers based on analysis of symmetry, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 17766–17771. Transparency document [6] F. Palmieri, Mitochondrial carrier proteins, FEBS Lett. 346 (1994) 48–54. [7] A. Tonazzi, N. Giangregorio, C. Indiveri, F. Palmieri, Identification by site-directed The Transparency document associated with this article can be mutagenesis and chemical modification of three vicinal cysteine residues inrat found, in online version. mitochondrial carnitine/acylcarnitine transporter, J. Biol. Chem. 280 (2005) 19607–19612. [8] N. Giangregorio, A. Tonazzi, L. Console, C. Indiveri, F. Palmieri, Site-directed mu- Funding tagenesis of charged amino acids of the human mitochondrial carnitine/acylcar- nitine carrier: insight into the molecular mechanism of transport, Biochimica et Biophysica Acta (BBA)-Bioenergetics, 1797 (2010) 839–845. This work was in part supported by PRIN (Progetti di Ricerca di [9] J.J. Ruprecht, M.S. King, T. Zögg, A.A. Aleksandrova, E. Pardon, P.G. Crichton, Interesse Nazionale) project n. 2017PAB8EM MIUR (Italian Ministry of J. Steyaert, E.R.S. Kunji, The molecular mechanism of transport by the mitochon- Instruction, University and Research) to CI, and in part by Programma drial ADP/ATP carrier, Cell 167 (2019) 435–447. [10] R. Springett, M.S. King, P.G. Crichton, E.R.S. Kunji, Modelling the free energy Operativo Nazionale project n. 01_00937 MIUR to CI. profile of the mitochondrial ADP/ATP carrier, Biochim. Biophys. Acta Bioenerg. 1858 (2017) 906–914. References [11] E.R. Kunji, A. Aleksandrova, M.S. King, H. Majd, V.L. Ashton, E. Cerson, R. Springett, M. Kibalchenko, S. Tavoulari, P.G. Crichton, J.J. Ruprecht, The transport mechanism of the mitochondrial ADP/ATP carrier, Biochim. Biophys. [1] F. Palmieri, M. Monné, Discoveries, metabolic roles and diseases of mitochondrial Acta 1863 (2016) 2379–2393. carriers: a review, Biochim. Biophys. Acta 1863 (2016) 2362–2378. [12] J.J. Ruprecht, A.M. Hellawell, M. Harding, P.G. Crichton, A.J. McCoy, E.R. Kunji, [2] E. Pebay-Peyroula, C. Dahout-Gonzalez, R. Kahn, V. Trézéguet, G.J. Lauquin, Structures of yeast mitochondrial ADP/ATP carriers support a domain-based al- G. Brandolin, Structure of mitochondrial ADP/ATP carrier in complex with car- ternating-access transport mechanism, Proc. Natl. Acad. Sci. U. S. A. 111 (2014) boxyatractyloside, Nature 426 (2003) 39–44. E426–E434. [3] E.R. Kunji, A.J. Robinson, The conserved substrate binding site of mitochondrial [13] M. Monné, F. Palmieri, E.R. Kunji, The substrate specificity of mitochondrial

715 N. Giangregorio, et al. BBA - Bioenergetics 1860 (2019) 708–716

carriers: mutagenesis revisited, Mol. Membr. Biol. 30 (2013) 149–159. proteins: purification, reconstitution, and transport studies, Methods Enzymol. 260 [14] N. Giangregorio, A. Tonazzi, L. Console, M. Galluccio, V. Porcelli, C. Indiveri, (1995) 349–369. Structure/function relationships of the human mitochondrial ornithine/citrulline [28] N. Giangregorio, A. Tonazzi, C. Indiveri, F. Palmieri, Conformation-dependent ac- carrier by Cys site-directed mutagenesis. Relevance to mercury toxicity, Int. J. Biol. cessibility of Cys-136 and Cys-155 of the mitochondrial rat carnitine/acylcarnitine Macromol. 120 (2018) 93–99. carrier to membrane-impermeable SH reagents, Biochimica et Biophysica Acta [15] C. Indiveri, V. Iacobazzi, A. Tonazzi, N. Giangregorio, V. Infantino, P. Convertini, (BBA)-Bioenergetics, 1767 (2007) 1331–1339. L. Console, F. Palmieri, The mitochondrial carnitine/acylcarnitine carrier: function, [29] A. Tonazzi, N. Giangregorio, C. Indiveri, F. Palmieri, Site-directed mutagenesis of structure and physiopathology, Mol. Asp. Med. 32 (2011) 223–233. the His residues of the rat mitochondrial carnitine/acylcarnitine carrier: implica- [16] N. Giangregorio, L. Console, A. Tonazzi, F. Palmieri, C. Indiveri, Identification of tions for the role of His-29 in the transport pathway, Biochimica et Biophysica Acta amino acid residues underlying the antiport mechanism of the mitochondrial car- (BBA)-Bioenergetics, 1787 (2009) 1009–1015. nitine/acylcarnitine carrier by site-directed mutagenesis and chemical labeling, [30] G.M. Loudon, D.E. Koshland, The chemistry of a reporter group: 2-hydroxy-5-ni- Biochemistry 53 (2014) 6924–6933. trobenzyl bromide, J. Biol. Chem. 245 (1970) 2247–2254. [17] L. Tang, L. Bai, W.H. Wang, T. Jiang, Crystal structure of the carnitine transporter [31] A. Shah, M. Aatif, M. Priyadarshini, M.S. Khan, F. Amin, B. Bano, Studies on the and insights into the antiport mechanism, Nat. Struct. Mol. Biol. 17 (2010) chemical modification of goat liver cystatin and the effect on its anti-papain in- 492–496. hibitory activity, J. Fluoresc. 22 (2012) 1627–1632. [18] D. Drew, O. Boudker, Shared molecular mechanisms of membrane transporters, [32] C. Indiveri, A. Tonazzi, N. Giangregorio, F. Palmieri, Probing the active site of the Annu. Rev. Biochem. 85 (2016) 543–572. reconstituted carnitine carrier from rat liver mitochondria with sulfhydryl reagents, [19] L. Napolitano, M. Galluccio, M. Scalise, C. Parravicini, L. Palazzolo, I. Eberini, Eur. J. Biochem. 228 (1995) 271–278. C. Indiveri, Novel insights into the transport mechanism of the human amino acid [33] F. Palmieri, C.L. Pierri, Structure and function of mitochondrial carriers - role of the transporter LAT1 (SLC7A5). Probing critical residues for substrate translocation, transmembrane helix P and G residues in the gating and transport mechanism, FEBS Biochim. Biophys. Acta Gen. Subj. 1861 (2017) 727–736. Lett. 584 (2010) 1931–1939. [20] X. Gao, L. Zhou, X. Jiao, F. Lu, C. Yan, X. Zeng, J. Wang, Y. Shi, Mechanism of [34] E.R. Kunji, M. Harding, Projection structure of the atractyloside-inhibited mi- substrate recognition and transport by an amino acid antiporter, Nature 463 (2010) tochondrial ADP/ATP carrier of Saccharomyces cerevisiae, J. Biol. Chem. 278 828–832. (2003) 36985–36988. [21] C.J. Law, P.C. Maloney, D.N. Wang, Ins and outs of major facilitator superfamily [35] S.V. Pande, R. Parvin, Carnitine-acylcarnitine translocase catalyzes an equilibrating antiporters, Annu. Rev. Microbiol. 62 (2008) 289–305. unidirectional transport as well, J. Biol. Chem. 255 (1980) 2994–3001. [22] C. Indiveri, V. Iacobazzi, N. Giangregorio, F. Palmieri, The mitochondrial carnitine [36] C. Indiveri, A. Tonazzi, F. Palmieri, Characterization of the unidirectional transport carrier protein: cDNA cloning, primary structure and comparison with other mi- of carnitine catalyzed by the reconstituted carnitine carrier from rat liver mi- tochondrial transport proteins, Biochem. J. 321 (1997) 713–719. tochondria, Biochim. Biophys. Acta 1069 (1991) 110–116. [23] C. Indiveri, N. Giangregorio, V. Iacobazzi, F. Palmieri, Site-directed mutagenesis [37] A.S. Mahadevi, G.N. Sastry, Cation-π interaction: its role and relevance in chem- and chemical modification of the six native cysteine residues of the rat mitochon- istry, biology, and material science, Chem. Rev. 113 (2013) 2100–2138. drial carnitine carrier: implications for the role of cysteine-136, Biochemistry 41 [38] S. Ressl, A.C. Terwisscha van Scheltinga, C. Vonrhein, V. Ott, C. Ziegler, Molecular (2002) 8649–8656. basis of transport and regulation in the Na(+)/betaine symporter BetP, Nature 458 [24] S.N. Ho, H.D. Hunt, R.M. Horton, J.K. Pullen, L.R. Pease, Site-directed mutagenesis (2009) 47–52. by overlap extension using the polymerase chain reaction, 77 (1989) 51–59. [39] A. Tonazzi, I. Eberini, C. Indiveri, Molecular mechanism of inhibition of the mi- [25] C. Indiveri, V. Iacobazzi, N. Giangregorio, F. Palmieri, Bacterial overexpression, tochondrial carnitine/acylcarnitine transporter by omeprazole revealed by proteo- purification, and reconstitution of the carnitine/acylcarnitine carrier from rat liver liposome assay, mutagenesis and bioinformatics, PLoS One 8 (2013) e82286. mitochondria, Biochem. Biophys. Res. Commun. 249 (1998) 589–594. [40] V. Porcelli, G. Fiermonte, A. Longo, F. Palmieri, The human gene SLC25A29, of [26] J.R. De Lucas, C. Indiveri, A. Tonazzi, P. Perez, N. Giangregorio, V. Iacobazzi, 25, encodes a mitochondrial transporter of basic amino acids, F. Palmieri, Functional characterization of residues within the carnitine/acylcar- J. Biol. Chem. 289 (2014) 13374–13384. nitine translocase RX2PANAAXF distinct motif, Mol. Membr. Biol. 25 (2008) [41] M. Monné, D.V. Miniero, L. Daddabbo, A.J. Robinson, E.R. Kunji, F. Palmieri, 152–163. Substrate specificity of the two mitochondrial ornithine carriers can be swapped by [27] F. Palmieri, C. Indiveri, F. Bisaccia, V. Iacobazzi, Mitochondrial metabolite carrier single mutation in substrate binding site, J. Biol. Chem. 287 (2012) 7925–7934.

716